Format

Send to

Choose Destination
Med Phys. 2014 Apr;41(4):041910. doi: 10.1118/1.4868457.

MR-based motion correction for PET imaging using wired active MR microcoils in simultaneous PET-MR: phantom study.

Author information

1
Center for Advanced Medical Imaging Sciences, Division of Nuclear Medicine and Molecular Imaging, Department of Imaging, Massachusetts General Hospital, Boston, Massachusetts 02114 and Department of Radiology, Harvard Medical School, Boston, Massachusetts 02115.
2
Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, Massachusetts 02129 and Department of Radiology, Harvard Medical School, Boston, Massachusetts 02115.
3
Center for Advanced Medical Imaging Sciences, Division of Nuclear Medicine and Molecular Imaging, Department of Imaging, Massachusetts General Hospital, Boston, Massachusetts 02114.

Abstract

PURPOSE:

Artifacts caused by head motion present a major challenge in brain positron emission tomography (PET) imaging. The authors investigated the feasibility of using wired active MR microcoils to track head motion and incorporate the measured rigid motion fields into iterative PET reconstruction.

METHODS:

Several wired active MR microcoils and a dedicated MR coil-tracking sequence were developed. The microcoils were attached to the outer surface of an anthropomorphic(18)F-filled Hoffman phantom to mimic a brain PET scan. Complex rotation/translation motion of the phantom was induced by a balloon, which was connected to a ventilator. PET list-mode and MR tracking data were acquired simultaneously on a PET-MR scanner. The acquired dynamic PET data were reconstructed iteratively with and without motion correction. Additionally, static phantom data were acquired and used as the gold standard.

RESULTS:

Motion artifacts in PET images were effectively removed by wired active MR microcoil based motion correction. Motion correction yielded an activity concentration bias ranging from -0.6% to 3.4% as compared to a bias ranging from -25.0% to 16.6% if no motion correction was applied. The contrast recovery values were improved by 37%-156% with motion correction as compared to no motion correction. The image correlation (mean ± standard deviation) between the motion corrected (uncorrected) images of 20 independent noise realizations and static reference was R(2) = 0.978 ± 0.007 (0.588 ± 0.010, respectively).

CONCLUSIONS:

Wired active MR microcoil based motion correction significantly improves brain PET quantitative accuracy and image contrast.

PMID:
24694141
PMCID:
PMC3978416
DOI:
10.1118/1.4868457
[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Wiley Icon for PubMed Central
Loading ...
Support Center